Methods
We selected 35 patient-derived metastatic lesion organoids (PDOs) from 18 pretreated and 17 chemonaive mCRC patients. We investigated transcriptomic and genomic changes related to acquired drug resistance, by combining drug sensitivity assessments with RNA and whole genome sequencing.
Results
PDOs from pretreated patients exhibited significantly higher tumor mutational load, more structural variants, and more deep deletions in driver genes at common fragile sites. Chemotherapy-related mutational signatures SBS17A, SBS17B and DBS5 were observed in PDOs from pretreated patients and the magnitude of these signatures correlated with previous exposure to 5-FU and oxaliplatin. PDOs derived from oxaliplatin-resistant patients captured the acquired resistance and did not show specific genomic alterations. PDOs derived from PDOs from oxaliplatin-resistant patients showed upregulation of ZNF300 and TGM2, along with enrichment of genes in the Hedgehog signaling pathway. In contrast, acquired resistance to 5-FU and irinotecan was only captured partially in PDOs. Acquired intrinsic resistance to irinotecan in pretreated PDOs correlated with specific genomic alterations: SBS17A and deep deletions in genes located at common fragile sites. These deletions were associated with a distinct gene expression profile with enrichment of BIRC3 and FBXO44. We identified 103 genes whose expression levels correlated with chemotherapy resistance in PDO drug screens. While oxaliplatin resistance correlated with gene expression in both primary and acquired resistant PDOs, correlations for SN-38 and 5-FU resistance were only observed in acquired resistant PDOs.
Conclusion
PDOs capture how chemotherapy induces accumulation of genomic mutations and phenotypic changes, which are related to exposure level and resistance. Acquired oxaliplatin resistance is phenocopied in PDOs by static tumor cell-intrinsic properties. In contrast, acquired resistance to irinotecan and 5-FU are only partially captured in PDOs, leaving a greater role for dynamic, transient cell states or tumor cell extrinsic factors. This study sheds light on the tumor biology of heavily pretreated patients and the value of pretreated models in unraveling resistance mechanisms.
